基于连续隐马尔可夫模型的仿真模型验证

摘要

A method based on continuous hidden Markov model (CHMM) is applied in the research on the simulation model validation. Firstly, the output feature vector of each selectuble simulation model is extracted in some segments. By using segmental K-means algorithm, CHMM for each simulation model is established and trained, the CHMM-base is built. Once the feature vector sequences of the real system are offered to CHMM-base, the probability outputs are generated. The validity of each selectable simulation model relative to the real system could be judged from the corresponding probability output. The method can be applied to model validation of aircraft guidance and control systems. The results show the reasonability and validity of the proposed method.%将连续隐马尔科夫模型(CHMM)应用于仿真模型验证研究,在分段提取各可选仿真模型输出特征向量序列的基础上,采用Segmental K-Means算法训练并建立各可选仿真模型的CHMM,进而构建模型库；将实际系统的特征向量序列作用于模型库,依据其概率输出判别各可选仿真模型相对于实际系统的有效性.最后通过实例分析,证明了该方法的有效性.